The long-distance movements made by humans through history are quickly erased by time but can be reconstructed by studying the genetic make-up of organisms that travelled with them. The phylogeography of the western house mouse (Mus musculus domesticus), whose current widespread distribution around the world has been caused directly by the movements of (primarily) European people, has proved particularly informative in a series of recent studies. The geographic distributions of genetic lineages in this commensal have been linked to the Iron Age movements within the Mediterranean region and Western Europe, the extensive maritime activities of the Vikings in the 9th to 11th centuries, and the colonisation of distant landmasses and islands by the Western European nations starting in the 15th century. We review here recent insights into human history based on phylogeographic studies of mice and other species that have travelled with humans, and discuss how emerging genomic methodologies will increase the precision of these inferences.
Mus musculus domesticus;
Co-phylogeography of humans and their close associates
Genetic markers can be used to track the colonisation and demographic history of organisms by fitting their genealogies to geographical and temporal frameworks (Box 1). This approach is known as phylogeography and, in the case of more than one organism, ‘comparative’ phylogeography can be used to infer common histories between them . Co-phylogeography, a special case of comparative phylogeography, was first used to describe the intimate linkage between the colonisation and demographic history of an organism and its parasites . Here we extend the term to encompass a wider suite of organisms that, through their close association with humans, share a common colonisation history, including organisms that humans exploit (e.g., plant and animal domesticates) as well as those that exploit humans (e.g., pathogens, parasites, pests). It is by moving these organisms around that humans create a mutual colonisation history.
Box 1. Phylogeographic analysis
To understand the colonisation history of specific geographical areas through phylogeography, the species of interest needs to be sampled both in the area it has colonised and in the possible source areas, and genetic analyses carried out within this geographic context. Over the years, phylogeographical studies have overwhelmingly been undertaken with a single type of molecular marker, mitochondrial DNA sequences. Genealogies can be constructed to examine relationships of such sequences and to make inferences about the colonisation history of the species in question. The standard way to present the resulting data is as a phylogenetic tree, often generated by Bayesian methods ; however, networks are also a valuable way to display the data  and an example is presented here (Figure I). Networks have the advantage of showing the relative frequencies of different sequences and can reflect alternative phylogenetic solutions. Figure I presents a situation where colonisation is not from the nearest possible source, as will often be the case with human-mediated colonisation, where human movements for various geographical and cultural reasons may not always involve the shortest distances. With a known substitution rate, which needs to be estimated with care , we can surmise when the colonisation event occurred. Thus, with relatively simple data, important aspects of colonisation history can be inferred. With good sampling, even using a single genetic marker, demographic analyses are also possible, for example, with Bayesian estimates of changes in population size over time , again dependent on an accurate substitution rate (see ). Sequences from multiple genes or genomic data allow more sophisticated analyses 70 and 71. Inferences about colonisation history also become much stronger if dated historical samples are available. The degraded DNA associated with such historical material can carefully be processed to yield genetic information using ancient DNA methodologies .
Figure I. Hypothetical network genealogy of a species found on three landmasses. The network distinguishes between two possible source areas (X and Y) in the colonisation of a third landmass (Z). The squares represent the relative positions of the different landmasses, colour-matched with circles in the network reflecting unique DNA sequences or ‘haplotypes’ of individuals from those landmasses. The sizes of the circles in the network are proportional to the frequency of the haplotypes. Each line indicates a single mutational step and bars on a line indicate a larger number of mutations. The presence in one of the possible source areas (X) of sequences that are identical or highly similar to those found in Z indicates that the colonisation was from X. The network displays a star-like phylogeny of haplotypes from Z; this is to be expected when there has been recent colonisation of an area by a small number of individuals with the same or closely related sequences. The ‘star’ consists of the common central haplotypes that were introduced and haplotypes that have arisen in the colonised area by one or a few mutational steps.
When genetic analyses of human-associated species do reveal a common history, they can be viewed as historical ‘proxies’ , and the phylogeography of these associates may fill in gaps in our knowledge and understanding of human movements and demographic history. It is this potential use of other organisms to tell us about human history that we explore here. We first develop the principles behind the use of ‘bioproxies’ as archaeological tools, and survey examples of them, before examining the house mouse (Mus musculus) as a detailed case-study.
Bioproxies in historical inference
Most of our knowledge about the human past comes from documentary evidence, analysis of material artefacts (Figure 1) and other vestiges of a human lifestyle, and interpretation of human remains. More recently, important insights into the colonisation and demographic histories of humans have been gained from genetic (phylogeographic) studies of both contemporary humans (using modern DNA) and archaeological human remains (using ancient DNA), such as the initial human colonisation of the Americas  and the recent colonisation of Scandinavia .
Figure 1. A Norse decorated comb of Viking Age from Mývatnssveit, Iceland, as an example of a material artefact providing evidence of human colonisation. This comb is made from deer antler (not available in Iceland) and bone, and was probably imported by the first settlers coming from Scandinavia or the British Isles (T.H. McGovern, personal communication).
Analogous to material artefacts, bioproxies may be considered ‘living artefacts’ – organisms taken to a place and left there by people moving between locations – artefacts with a (potentially enormous) information content dependent on the depth of analysis of the genomes of the organisms. Bioproxy phylogeography may be more valuable than human phylogeography in determining the human colonisation history of an area, for instance in situations where the human colonists and their descendants have left or died out, or if the genetic signal of the original human founders is obscured by subsequent immigration. Bioproxies also avoid ethical concerns of working directly on humans.
The phylogeography of bioproxies may be an alternative or complement to the phylogeography of humans when considering any period from the origin of humans onwards. However, bioproxies become particularly valuable at elucidating human movements after the formation of settled societies, agriculture, and transport . This is when people became closely associated with a large number of different types of organisms and developed an increasingly complex colonisation history brought about by recurrent immigration and emigration, which bioproxies can help to unravel.
Categories of bioproxies
Bioproxies can be divided into three main categories, representing many species used in phylogeographic studies: domesticates (and other purposefully transported organisms), commensals (unintentionally transported organisms), and pathogens and parasites.
Numerous phylogeographic studies on modern and archaeological specimens have investigated the domestication centres for different species and the spread of those domesticates thereafter (e.g., sheep Ovis orientalis, cattle Bos primigenius8 and 9, grape Vitis vinifera, and maize Zea mays). These studies reveal not only aspects of the spread of agriculture, but also human migrations by people who carried domesticates with them. This is exemplified by studies on the pig (Sus scrofa), examining the spread of the Neolithic through Europe , and human colonisation of Oceania . Care is needed, however, in interpreting genetic data on extant breeds and cultivars , because replacement events (humans bringing in new breeds) may wipe out earlier colonisation events, although original colonists may survive in a feral state . Domesticates are not the only exploited species that may act as bioproxies. For example, genetic studies (mitochondrial and nuclear DNA sequences) of the snail Tudorella sulcata indicate that living individuals were moved by people between Sardinia, France, and Algeria . One likely explanation is that the shell of this snail was used as an ornament in the Neolithic period, making it a good bioproxy for Neolithic movements in the Mediterranean.
The term ‘commensal’ has various usages, but here describes a spectrum of species that opportunistically exploit habitats and food sources around humans (particularly dwellings and nearby areas) and are readily transported by humans. Commensals in this sense can vary between having no discernible negative effect on humans to being pests, and the same organism can span this continuum under different circumstances. Interestingly, several commensals are important biological models, most notably the fruit fly , rat , and mouse (the focus of this article); presumably their commensalism makes them noticeable and available to humans, and likely to fare well in captivity. The progenitors of the laboratory fruit fly and rat, Drosophila melanogaster and Rattus norvegicus, have predominantly been moved around by people in the past 200–300 years 16 and 18, and thus have limited value as bioproxies for human history. However, other commensal rats, including the ship rat, R. rattus (Box 219 and 20), and in particular the Pacific rat, R. exulans, have been of more interest in this regard. Although it lives around human dwellings, as expected for a typical commensal, R. exulans, similarly to the pig, may actually have been transported deliberately as food by Polynesians . However, phylogeographic data (mitochondrial DNA) on ancient and modern specimens of the pig and R. exulans provide different perspectives: the former fit the classical model for the colonisation of Oceania by a single spread of the Lapita people , and the latter fit a model of multiple introductions and therefore multiple waves of colonisation by people 3 and 22. The value of using multiple bioproxies of various categories to examine a specific problem of human colonisation history is illustrated further in Box 2. As with other types of evidence relating to human history, conflicting results may be obtained, and there is an advantage in having data from a variety of systems.
Box 2. Co-phylogeography and the Madagascar conundrum
Madagascar lies 400 km from the coast of East Africa, but the most common language, Malagasy, is an Austronesian language closely related to Barito of Borneo , which is over 7500 km away. The earliest occupation site on the island dates to the mid-1st millennium AD, but the exact origin, timing, and route taken by these early immigrants remains unknown . If Austronesians crossed the open ocean on a direct voyage from Southeast Asia to Madagascar, it would require a major revision of our understanding of early maritime capabilities (Figure I).
Figure I. A dhow. The dhow has been used for centuries in the Indian Ocean region, often sailing with the monsoon winds, and was used in early contacts between Madagascar, East Africa, and Arabia.
With a lack of archaeological or historical evidence, much of what we know about the settlers reaching Madagascar comes from commensal and domestic proxies. A study using autosomal microsatellites of cattle (Figure II) supports the hypothesis of a direct route of contact between Southeast Asia and Madagascar : there is substantial introgression of Asian zebu alleles into Malagasy cattle, unlike the majority of the rest of Africa which is characterised by predominantly African cattle genotypes (although East Africa also shows high levels of admixture). Also in support of the Southeast Asian link, a mitochondrial cytochrome b genealogy places the Malagasy pygmy shrew (Suncus madagascariensis) into the widespread Etruscan shrew complex (Suncus etruscus), meaning that it is likely to be a relatively recent introduction from South or Southeast Asia rather than a Malagasy endemic . Commensal phylogeographies of ship rats and house mice using mitochondrial markers (cytochrome b + D-loop and D-loop, respectively) reveal a connection between Malagasy individuals and those from the Arabian Peninsula, probably reflecting the Arab trade that occurred with Madagascar early in the 2nd millennium AD, or possibly through even earlier connections 20 and 28.
Figure II. Indian Ocean bioproxies: zebu, taro, and house gecko (Hemidactylus frenatus).
Rice (Oryza sativa), the greater yam (Dioscorea alata), and the coconut tree (Cocos nucifera) are cultivated plants of Madagascar that were early introductions from South or Southeast Asia . Bananas (Musa sp.) and taro (Figure II; Colocasia esculenta) are also of Asian origin, and were introduced into Africa and Madagascar probably through direct maritime exchanges because they are archaeologically absent from intermediate areas such as Arabia . These plants could potentially be highly informative bioproxies over various time-periods, as could commensal stowaways such as the house shrew (Suncus murinus) and house geckos (Figure II; Hemidactylus spp.).
Of the various possible human associates, pathogens and parasites (as well as other symbionts) are the proxies with the potential to show the closest match in terms of colonisation and demographic histories. This is vividly illustrated by the co-phylogeography of humans and the stomach-living bacterial pathogen Helicobacter pylori, reviewed in . The initial spread of humans out of Africa and the initial colonisation of Europe and the Americas by humans are revealed through multi-locus DNA sequencing of H. pylori. Much more recent events, such as secondary colonisation of the Americas by Europeans and Africans, can also be detected. The bacterial data also reflect human demographic features such as population bottlenecking. The information from genotyping H. pylori appears to be remarkably close to genotyping of humans themselves, and this is explained by the pathogen being readily spread within human families , effectively vertical transmission. Pathogens and parasites where transmission is more horizontal are expected to display phylogeographic patterns less closely matched to their human hosts. Nevertheless, there may still be a signal that matches human movements and colonisation history 24 and 25. For instance, the bacterial pathogen that leads to leprosy, Mycobacterium leprae, is one of several pathogens for which the phylogeography can be related to human movements along major trading routes, such as the Silk Road (linking Asia to Europe) and the slave trade (linking West Africa and the Americas) . One limitation of pathogens and parasites is that they are sampled from existing human populations and may not be of use in recovering data about extinct human populations, whereas other bioproxies able to persist independently of humans (e.g., ‘feral’ populations of domesticates or commensals) will still carry a signal.
The house mouse as a model system
The house mouse Mus musculus, the progenitor of the laboratory mouse, is a very familiar commensal living together with, and feeding on the food of, people and their livestock. This close association with humans first began with the M. m. domesticus subspecies around 12 000 years ago in the Near East  when mice exploited the niche offered by burgeoning human settlements and grain stores. They have accompanied humans with trade and transport ever since, reaching a near-global distribution . Through this widespread and long-lasting association with humans, mice offer extraordinary possibilities as bioproxies (cats can also be considered in the same co-phylogeographic grouping; Box 3).
Box 3. A game of cat-and-mouse
Over time, many of the species of felids worldwide have been tamed as pets and used for hunting and pest control . However, only the wildcat (Felis silvestris) has been domesticated, and a domestic form is found globally and has been the subject of detailed genetic analysis (e.g., ). Although cat domestication was once thought to have occurred in Ancient Egypt, archaeological data show that this took place considerably earlier , and microsatellite and mitochondrial DNA data indicate that the domestication occurred in the Near East . Thus, it is most reasonable to suggest that domestication started with the rise of agriculture in the Neolithic of the Near East when wildcats were probably attracted to human settlements by rodents that took advantage of stored food or waste dumps and, due to their ability to control rodent pests, it became beneficial to tame and selectively breed them 79 and 82. The substantial haplotype variation in mitochondrial DNA in domestic cats suggests a protracted domestication process . From the archaeological and historical record we present three examples showing that human, cat, and mouse history are bound together. First, the unearthing of human, cat, and house mouse remains in Cyprus from at least 9000 years ago, reflecting early Neolithic movement away from the Near East 27, 79 and 81. Second, the Viking Age occurrence of cat and mouse bones at Mývatnssveit, Iceland 38 and 83, the same area that yielded the material artefact in Figure 1 in main text. Third, from the Age of Discovery, there are written records that cats were deliberately taken on ships for control of rodents and spread worldwide as a result (e.g., ). Therefore, there is a clear potential to use domestic cats, as well as wild mice, as bioproxies for human colonisation and demographic history. To date there has been little phylogeographic work with molecular markers on domestic cats, although a microsatellite study indicates that West European settlers brought cats over to the New World . However, various intriguing suggestions are made in  on the history of dispersal of coat-colour variants and other morphological traits in cats, for example that the geographic distribution of the sex-linked orange coat colour variant may relate to Viking movements.
Current evidence suggests that M. musculus originated in the north of the Indian subcontinent where it differentiated into three main subspecies, domesticus, musculus, and castaneus, initially through geographic isolation in the Pleistocene 28 and 29; these subspecies independently adopted a commensal human niche. The subsequent range expansion of M. m. domesticus is discussed in more detail below. The M. m. musculus subspecies was spread as a human commensal to northern Eurasia, covering most of the landmass except Western Europe, but including colonisation of Japan 30 and 31. In a comparatively recent human-mediated expansion, M. m. castaneus moved further east and south into southern India, Southeast Asia and Japan, as well as Africa and the Indo-Pacific in historical times . Laboratory mice show varied contributions of these subspecies: the classical inbred strains are overwhelmingly M. m. domesticus but with a genetic contribution of Japanese mice, themselves hybrids of M. m. musculus and M. m. castaneus.
House mice are highly adaptable, reproduce rapidly and, in the absence of competitors, can readily establish new populations. It is seemingly rare for an invading female to successfully integrate into an existing population, whereas males can do so . This has made mice ideal proxies to study human history because the signature of the founding females will generally be maintained in maternally inherited mitochondrial DNA, and thus provides an indicator of early human exchanges. Subsequent human trade and migration linked to secondary colonisations of (generally male) mice can be inferred through nuclear (including Y-linked) markers 34 and 35. Large islands, archipelagos, and continental areas may have multiple primary colonisations that can be revealed with mitochondrial DNA, as exemplified by studies on the sub-Antarctic Kerguelen archipelago . Current studies of mitochondrial DNA benefit from a substantial database of D-loop (control region) sequences 32, 37 and 38.
A clear understanding of the phylogeography of house mice, their colonisation history, and the degree to which different lineages (including subspecies) come into secondary contact, underpins other fundamental evolutionary studies in which mice are used as a model, such as those related to speciation, adaptation, and genome evolution 39, 40 and 41.
Western house mice and human history
The phylogeography of the western subspecies of house mouse (M. m. domesticus) provides a case-study for the use of proxies in human history. The subspecies is found in Western Europe, Africa, Australasia, and the Americas, and the observed patterns of mitochondrial D-loop variation provide insights into human movements within and into those areas at various historical periods. Six D-loop clades with subclades within them have been named , using similar nomenclature to that in humans. Alternative phylogenetic methods have generated 11 haplogroups , many of which can be equated to subdivisions within the six clade scheme . Figure 2 shows the distribution of the six clades over a region from the Near East to the North Atlantic.
Figure 2. Phylogenetic tree of D-loop sequences of western house mice including six named clades, adapted from . The distribution of the six clades within the vicinity of Western Europe is shown ( and references within; 38, 49, 59 and 86; S.I. Gabriel et al., unpublished). All sample locations relate to modern specimens except one in Iceland (clade F) which relates to Viking Age specimens (marked with an asterisk).
The greatest genetic diversity (mitochondrial and X-linked) of the western house mouse occurs in the Near East, consistent with this area being the origin of its commensalism and expansion . Genetic studies on humans have been used to examine human expansion from the Near East , including movements into Western Europe during the Neolithic 44 and 45. The mouse expansion into Western Europe occurred much later (in the Iron Age, see below), and mice can therefore only be used as a proxy for Neolithic humans in the Near East region. For example, the colonisation of Anatolia from the Fertile Crescent by house mice during the Neolithic  may provide interesting clues in light of current interest in Anatolia for the spread of languages and agriculture further afield . Increasing amounts of D-loop data have been collected for mice in this area, where clades A, B, and C are particularly well-represented (Figure 2) 37 and 47.
Even though house mice occurred around human dwellings during the Neolithic in the Near East, it was not until the Iron Age, about 3000 years ago, that they expanded into Western Europe . The delay may have been caused by the small size of European settlements before the Iron Age, when any house mice that invaded would have been unable to outcompete the native wood mouse, Apodemus sylvaticus; only in larger Iron Age settlements with more commensal habitat could the house mice escape this competition . The increase in maritime transportation during the Iron Age, including Phoenician activities , accelerated the sweep of both mice and people through the Mediterranean basin and North Africa. This led to the spread of various mouse D-loop clades across the region 37, 42 and 48; the distribution of clade B is notably representative of this phase (Figure 2). The subsequent expansion of mice northwards  is also represented by genetic data, with the penetration of Mediterranean D-loop clades C and D into Central Europe (Figure 2) . Mice may not only reflect Iron Age linkages between the Mediterranean and central Europe, but also between Britain and nearby areas of continental Europe in an area marked by clade E (Figure 2 and Box 4) 42, 49 and 50.
Box 4. Unexpected phylogeographic findings in the house mouse
One of the most surprising findings of mouse phylogeography is a connection between the Madeira archipelago (off the Atlantic coast of North Africa) and the Danish Vikings. Madeira was officially discovered by Portuguese explorers in 1419 AD, and the vast majority of its contact ever since has been with mainland Portugal. However, despite intensive sampling, almost no D-loop sequences are shared between Madeira and Portugal, instead there is a striking similarity between the sequences on Madeira and in northern Europe, both belonging to clade D (see Figure 2 in main text) 28 and 87. This can be explained if 9th century Danish Viking boats with northern European mice were blown off course when travelling to the Mediterranean, landing on Madeira, and leaving mice as the only evidence of their visitation. In a recent development, mice on some of the islands of the Azores have been shown to have a strong representation of the clade linked to Norwegian Vikings (clade F; see Figure 2 in main text and above), again with virtually no similarity to mice from Portugal, the country of discovery and dominance (S.I. Gabriel et al., unpublished; ). Evidence from Medieval maps suggests to some historians that Norwegian Vikings could have visited the Azores , but the mouse phylogeographic data are the first independent findings in favour of that idea.
More speculatively, the mouse D-loop data may also track a long-distance trade connection involving the Phoenicians, already suggested as agents responsible for transporting house mice around the Mediterranean basin . Clade E is largely found in north-western continental Europe and the British Isles, but also in the Mediterranean including the east, albeit at low sampling levels (see Figure 2 in main text). From its presumptive eastern Mediterranean origins, no Clade E samples have been found over continental Europe even though it arrived in north-western Europe during the Iron Age 42 and 50. The Phoenicians ran a trading empire centred on the eastern and southern Mediterranean from the 2nd millennium BC, but it is likely that they also engaged in more long-distance exchange of goods .They may have transported Clade E by boat from the eastern Mediterranean to north-western Europe as a consequence of an early trading route for tin, a link for which there is currently no supporting archaeological evidence . The mouse data in support of this is still scanty but hints at exciting future possibilities using mice as bioproxies.
It is then interesting to reflect on the journey of the clade E mitochondrial haplotype now found in classic inbred laboratory strains 42, 50, 91 and 92. The mitochondrial lineage apparently originated in the eastern Mediterranean/Near East and was taken to Britain in the Iron Age, possibly by the Phoenicians. There, it can be suggested that mice with the relevant haplotype were domesticated in Britain, probably during the 19th century, and then transported as domestic ‘fancy’ mice no later than the early 20th century to the US and to Harvard to contribute to the formation of the classic inbred strains. Alternatively, after the mice were transported to Britain during the Iron Age, they were subsequently taken as wild individuals to the US and were domesticated there.
About a thousand years ago, the Norwegian and Danish Vikings had an important presence in the Atlantic periphery of Western Europe. They reached as far as Greenland and Newfoundland in the west and into the Mediterranean basin in the east . The Vikings undoubtedly transported mice , and the distributions of clades D and F in particular (Figure 2) appear to have been influenced by Viking dispersal. The value of mice as proxies for Viking history is explored further below and in Box 4.
Age of Discovery
The spread of house mice to coastal Western Europe during the Iron Age and Viking Age enabled a yet further expansion of the species. This area includes the countries that sent out ships that explored and subsequently settled much of the rest of the world from the early 15th century onwards, in other words Portugal, Spain, Britain, France, and The Netherlands. The ships used by the explorers and ensuing settlers would have been infested by mice (and rats), explaining why Captain Cook took cats on his vessels (Box 3) . The mouse genetic data currently available from countries settled by Europeans following the Age of Discovery have a signal that largely matches what is expected from our historical knowledge of the period. For instance, clades E and F, the best-represented clades in the British Isles (Figure 2), are predominant in mice from Australia and New Zealand 49 and 53. However, the mice from the Falkland Islands exhibit a more intriguing pattern. These islands are a British dependency in the South Atlantic, and some of the mice have clade E haplotypes, probably derived from Britain . There are also other mice carrying three D-loop haplotypes (from clade F) identical to those only otherwise found on two islands of the Azores archipelago (S.I. Gabriel et al., unpublished). It is probable that mice were taken from the Azores to the Falklands, and this could have been at the time of discovery rather than at the time of settlement; in the absence of competitors mice can thrive away from human settlements . Mice coming from the Azores may indicate Portuguese discovery, but the Azores could also have been a stopping point for British ships . Although inconclusive in this particular situation, house mice provide evidence of the first discovery of islands and landmasses by the Western European maritime powers, information of value to historians because of the secrecy often associated with such discoveries .
In-depth analysis of the co-phylogeography of house mice and Norwegian Vikings
During the Viking period, people from Norway embarked on extensive conquest, colonisation, and settlement of new areas from the late 8th century (Figure 3). Material artefacts reflect these linkages (Figure 1), but human genetics have also been informative. This has shown a general pattern of increasing Norwegian genetic ancestry in the Scottish Western Isles, Orkney, Shetland and Faroe, reaching a maximum in Iceland 56, 57 and 58, where an estimated 64% of the ancestry is Norwegian and 36% Gaelic . A remarkably similar colonisation pattern is seen in the distribution of house mouse mitochondrial DNA, where clade F is restricted almost exclusively to areas under Norwegian Viking influence (Figure 2 and Figure 3) 38, 42 and 50. Using ancient DNA from Viking Age house mouse bones, this clade (indeed, the same haplotype) was also found from the early colonisation period in Iceland continuously to the present day . Mouse samples from the Viking period in Greenland (which lasted until the mid to late 15th century, when the human settlements were abandoned) all belonged to a haplotype one mutation distinct from the common Icelandic haplotype. Like the Vikings themselves, the old Greenlandic strain apparently died out and modern mice sampled in Greenland reflect the recent Danish settlement there (clade D). No trace of clade F was found in modern Newfoundland mice (no ancient samples were available), such that the very short-lived Norwegian Viking colony there (in the early 11th century) left no detectable genetic signature in the mice; again the mice reflect recent human colonisation (probably from southern Britain: clade E). The human patterns of successful colonisation (Iceland) or colonisation and subsequent failure (Greenland, Newfoundland) are mirrored in house mice.
Figure 3. Maps of the North Atlantic region, showing (top) the distributions of the house mouse D-loop clades (circles coloured as in Figure 2 and citations also as in that figure) and (bottom) the Norwegian Viking Kingdom around the 10th century (stylised) with approximate routes of Viking colonisation. Clade F (blue) appears to track Norwegian Vikings between western Norway, the Scottish Isles, Ireland, Iceland, and Viking Age Greenland, and clade D (red) reflects the movement of house mice from southern Norway to the Faroe Islands at the same period (note that the modern sample from Greenland (clade D) is probably a modern introduction from Denmark because it belongs to the subspecies Mus musculus musculus but carries Mus musculus domesticus mitochondrial DNA.) The clade E mice from Newfoundland appear to be the result of European colonisation from Britain much later than the Vikings. Asterisks show localities where DNA has been obtained from Viking Age mouse bones; all were clade F (blue).
Here, we see a very clear co-phylogeographic pattern of house mice and humans (of European descent), and the indisputable potential of house mice as bioproxies for Norwegian Viking history. However, the situation is not completely clear-cut. Norway itself is occupied by a mix of different mouse clades (Figure 3), potentially explained as the simultaneous multiple arrivals of mice from various areas via maritime routes, reflecting the sudden explosion of seafaring expeditions from Norway during the Viking period . The Faroe Islands are occupied by clade D which may have arrived from southern rather than northern Norway (Figure 3) , consistent with what is known historically about the human colonisation of the islands .
Overwhelmingly, phylogeographic studies on the western house mouse have been conducted with a single mitochondrial marker, the D-loop. However, greater precision can be obtained from genomic approaches. Given that maternally inherited markers may be particularly valuable to yield information on first colonisation of an area, complete mitochondrial genomes are an obvious and relatively easy first step . However, the house mouse as a model will not properly be exploited until use is made of tools to investigate the nuclear genome, such as the 623 124 SNP Mouse Diversity Array . These will allow analysis of multiple colonisations of a particular area, and there are also likely to be interesting patterns involving selection and adaptive introgression . Genomic data of this sort are amenable to demographic modelling, including the Approximate Bayesian Computation approach . Figure 4 shows an example of a hypothetical population history of house mouse populations involved in colonisation events. With sufficient depth and quantity of genomic data, it would be possible to reveal scenarios such as this with demographic modelling. Because much of the house mouse history is in synchrony with human populations, if genetic analysis can determine demographic changes of this type in mouse populations over time, the mouse can be a particularly valuable proxy. What is true for the house mouse can also be applied to other bioproxies, and for any of these types of study there is a substantial benefit to include specimens contemporary to the actual events being modelled (via ancient DNA). In this way the full power of genomes and genetics can make a substantial difference in historical inference. Such studies will not replace evidence from material artefacts and documentary sources, but they have the potential to have a very large impact on our understanding of history.
Figure 4. Hypothetical house mouse demographic scenarios resulting from human activities. The thick black line represents the size of the house mouse population as it changes through time. Expansion of humans into new ranges (for example, onto islands, shown as transport by boats) will lead to a split in the house mouse population, while human demographic expansion or change in agricultural practice favouring mice will lead to a house mouse population expansion. Human population contractions or pest control will lead to a decline in the house mouse population. Disappearance of humans from an area may lead to local extinction of the mice, as has been suggested for Greenland  and is known to have happened in St. Kilda 42 and 93.
Studying humans is the most direct route to understanding ourselves, and just as biomedical studies on mice and other species can help us understand human physiology and genetics, so studies on bioproxies can contribute substantially to our understanding of human colonisation and demographic history. Many organisms can potentially be used as bioproxies (e.g., domesticates, commensals, pathogens, and parasites) and, by combining these in a co-phylogeographic framework with human data, we can use them to study human populations during particular periods of historical interest. House mice have lived with and been moved around by people for ∼12 000 years, and there are many instances where studies on their colonisation history are linked with human phylogeographic studies, such as the Neolithic expansion through Europe , the Phoenician spread through the Mediterranean , and the Viking colonisation of North Atlantic islands 56, 57 and 58. So far, studies on house mice have largely used mitochondrial D-loop sequences. In the future, applying the combination of genomic methods and demographic modelling to house mice will allow much more precise understanding of particular historical events and the complexity of repeated human (and therefore murine) colonisations into one geographical area. Typically treated as vermin, house mice could rather be considered as genetic historical treasures that are, in our collective experience, a numerically fast-declining part of our cultural heritage in many countries (in Western Europe at least).
We thank Jeremy Herman, Angela Douglas, members of the Searle laboratory, and anonymous referees for comments on the manuscript. We are very grateful to the following for photographs: Thomas McGovern (Viking comb), Luke Kirkwood (dhow), Ania Kotarba-Morley/Alison Crowther (zebu), Ilaria Grimaldi (taro), and Anthony Cheke (gecko). E.P.J. gratefully acknowledges a BIOCONSUS fellowship (Mammal Research Institute PAS), H.M.E. is funded by a European Research Council Grant (206148) through the Sealinks Project and S.I.G. acknowledges support from the Fundação para a Ciência e a Tecnologia (Portugal).
Note to users: Corrected proofs are Articles in Press that contain the authors' corrections. Final citation details, e.g., volume/issue number, publication year and page numbers, still need to be added and the text might change before final publication.
Although corrected proofs do not have all bibliographic details available yet, they can already be cited using the year of online publication and the DOI , as follows: author(s), article title, journal (year), DOI. Please consult the journal's reference style for the exact appearance of these elements, abbreviation of journal names and use of punctuation.
When the final article is assigned to an issue of the journal, the Article in Press version will be removed and the final version will appear in the associated published issue of the journal. The date the article was first made available online will be carried over.